optimization/8156: -mcpu=pentium4 generates wrong fp results in glibc-2.3

[Andreas Steinmetz]

The following reply was made to PR optimization/8156; it has been noted by GNATS.

From: Andreas Steinmetz <ast@domdv.de>
To: gcc-gnats@gcc.gnu.org,  gcc-prs@gcc.gnu.org,  bregor@sf.anu.edu.au, 
 gcc-bugs@gcc.gnu.org,  nobody@gcc.gnu.org
Cc:  
Subject: Re: optimization/8156: -mcpu=pentium4 generates wrong fp results
 in glibc-2.3
Date: Mon, 07 Oct 2002 13:13:19 +0200

  > This is a glibc bug, not a gcc bug.
 
 I'm not so sure that it is a problem of glibc only. In my case building 
 glibc 2.3 with gcc 3.2 (release)/binutils 2.13 (release) succeeds in all 
 cases but:
 
 1. CFLAGS="-O3 -mcpu=pentiumpro -march=pentiumpro" produces a working
     glibc
 
 2. CFLAGS="-O3 -fomit-frame-pointer -mcpu=pentiumpro -march=pentiumpro"
     produces a glibc with defunct pthreads (hangs and segfaults of
     threaded processes), strace-ing these processes indicate defunct
     signal handling
 
 As it seems that "-fomit-frame-pointer" does trigger the problems it 
 could well be that there is a gcc bug.